Kazakov, Viktor;
(2025)
Machine Learning Applications in
Fixed Income Markets and
Correlation Forecasting.
Masters thesis (M.Phil), UCL (University College London).
Preview |
Text
MPhil_corrected.pdf - Accepted Version Download (1MB) | Preview |
Abstract
In this work, we present two applications of classic machine learning algorithms in fixed income markets and correlation forecasting. We begin by developing a supervised learning model for the selection of bonds based on their relative value, trained on cross-sectional “static” data. We then expand this analysis to the dynamic setting by presenting a time series directional change/directional forecast framework (DC/DF). We apply the DC/DF framework to paired time series of rolling correlation between asset classes, with the objective of forecasting changes in the correlation of the asset classes at “intrinsic time” intervals. Forecasting changes in the correlation between asset classes can be applied in contexts such as portfolio optimization, risk management, and trading. The proposed frameworks are built on solid statistical fundamentals. We justify the choice of models by conducting two large-scale benchmarking studies to determine which algorithms tend to be most accurate, on average. We then propose a statistical framework for comparing the results of the machine learning models. This framework rests on a methodology for building confidence intervals around point estimates and non-parametric statistical tests for model comparison. When combined, the two models form a comprehensive framework for bottom-up securities selection and top-down allocation between different asset classes. The two frameworks are complementary and have direct practical use cases for value investing and portfolio diversification strategies. To the best of our knowledge, the proposed model validation methodology has not been previously applied to specifically solve financial problems. Furthermore, this is the first work that applies the proposed directional change/directional forecasting framework to time series of rolling correlation.
| Type: | Thesis (Masters) |
|---|---|
| Qualification: | M.Phil |
| Title: | Machine Learning Applications in Fixed Income Markets and Correlation Forecasting |
| Open access status: | An open access version is available from UCL Discovery |
| Language: | English |
| Additional information: | Copyright © The Author 2025. Original content in this thesis is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) Licence (https://creativecommons.org/licenses/by-nc/4.0/). Any third-party copyright material present remains the property of its respective owner(s) and is licensed under its existing terms. Access may initially be restricted at the author’s request. |
| UCL classification: | UCL UCL > Provost and Vice Provost Offices > UCL BEAMS UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Engineering Science > Dept of Civil, Environ and Geomatic Eng |
| URI: | https://discovery.ucl.ac.uk/id/eprint/10207422 |
Archive Staff Only
 |
View Item |
